Illumination Invariant Face Tracking on Smart Phones using Skin Locus based CAMSHIFT

Bui, Hoang Nam,Kim, SooHyung,Na, In Seop THE KOREAN INSTITUTE OF SMART MEDIA 2013 스마트미디어저널 Vol.2 No.4

This paper gives a review on three illumination issues of face tracking on smart phones: dark scenes, sudden lighting change and backlit effect. First, we propose a fast and robust face tracking method utilizing continuous adaptive mean shift algorithm (CAMSHIFT) and CbCr skin locus. Initially, the skin locus obtained from training video data. After that, a modified CAMSHIFT version based on the skin locus is accordingly provided. Second, we suggest an enhancement method to increase the chance of detecting faces, an important initialization step for face tracking, under dark illumination. The proposed method works comparably with traditional CAMSHIFT or particle filter, and outperforms these methods when dealing with our public video data with the three illumination issues mentioned above.

REGULAR VARIATION AND STABILITY OF RANDOM MEASURES

Nam Bui Quang,Phuc Ho Dang 대한수학회 2017 대한수학회지 Vol.54 No.3

The paper presents a characterization of stable random measures, giving a canonical form of their Laplace transform. Domain of attraction of stable random measures is concerned in a theorem showing that a random measure belongs to domain of attraction of any stable random measures if and only if it varies regularly at infinity.

Rhodanobacter soli sp. nov., isolated from soil of a ginseng field.

Bui, Thi Phuong Nam,Kim, Yeon-Ju,Kim, Hobin,Yang, Deok-Chun Society for General Microbiology 2010 International journal of systematic and evolutiona Vol.60 No.12

<P>Strain DCY45(T) was isolated from soil of a ginseng field in Pocheon Province, Korea. Strain DCY45(T) was Gram-negative, oxidase- and catalase-positive, motile and rod-shaped and produced yellow pigments on R2A agar. The organism grew optimally at 30 C and at pH 7.0. The G+C content of the genomic DNA was 65.4 mol%. The predominant respiratory quinone was Q-8. The major fatty acids were iso-C(17?:?1)ω9c, iso-C(16?:?0) and iso-C(15?:?0). Phylogenetic analysis based on the 16S rRNA gene sequence was used to determine the taxonomic position of strain DCY45(T), which is most closely related to species of the genus Rhodanobacter, with similarity levels of 96.0-98.4?%; DNA-DNA relatedness with related strains was lower than 60?%. Strain DCY45(T) differed significantly from related type strains in phenotypic characteristics. On the basis of these phenotypic, genotypic and chemotaxonomic studies, strain DCY45(T) represents a novel species of the genus Rhodanobacter, for which the name Rhodanobacter soli sp. nov. is proposed. The type strain is DCY45(T) (=KCTC 22620(T) =JCM 16126(T)).</P>

Gradient-Flow Tensor Divergence Feature for Human Action Recognition

BUI, Ngoc Nam,KIM, Jin Young,KIM, Hyoung-Gook 'Institute of Electronics, Information and Communi 2016 IEICE transactions on fundamentals of electronics, Vol.ea99 No.1

<P>Current research trends in computer vision have tended towards achieving the goal of recognizing human action, due to the potential utility of such recognition in various applications. Among many potential approaches, an approach involving Gaussian Mixture Model (GMM) supervectors with a Support Vector Machine (SVM) and a nonlinear GMM KL kernel has been proven to yield improved performance for recognizing human activities. In this study, based on tensor analysis, we develop and exploit an extended class of action features that we refer to as gradient-flow tensor divergence. The proposed method has shown a best recognition rate of 96.3% for a KTH dataset, and reduced processing time.</P>

Internet of agents framework for connected vehicles: A case study on distributed traffic control system

Bui, Khac-Hoai Nam,Jung, Jason J. Elsevier 2018 Journal of parallel and distributed computing Vol.116 No.-

<P><B>Abstract</B></P> <P>This study focuses on the distributed traffic control system by inspiration of advanced connected vehicle technology. In this regard, we introduce an Internet of Agents (IoA) framework for connected vehicles where agents make their own decisions to improve the effectiveness of the system by connectivity and automatic negotiation with other agents. Specifically, each of the connected vehicles can be regarded as an agent which is able to communicate and collaborate with others based on Vehicle-to-Vehicle (V2V) communication technologies. A case study on distributed traffic control system without traffic signal is presented in this paper. In particular, we consider traffic control at intersection problem as a group mutual exclusion problem where only connected vehicles in non-conflict relationship are able to enter the core of intersection simultaneously. Therefore, we extend the Ricart–Agrawala based-logical clock algorithm to deal with this problem. Various parameters (e.g., number of message exchange, average waiting time, total number of vehicle passing) have been measured to evaluate our approach. The simulated results show that our approach outperforms compared with existing traffic systems and previous works.</P> <P><B>Highlights</B></P> <P> <UL> <LI> A new concept “Internet of Agents” is proposed to support agent collaboration on IoT environments. </LI> <LI> We introduce a novel framework for traffic control without traffic signal systems that are not required for self-driving cars. </LI> <LI> As a case study, the proposed method has been applied to distributed traffic control. </LI> </UL> </P>

Micro-Bubble Flotation of Freshwater Algae: A Comparative Study of Differing Shapes and Sizes

Bui, Thi Thuy,Nam, Seong-Nam,Han, Mooyoung M. Dekker 2015 Separation science and technology Vol.50 No.7

Cooperative game-theoretic approach to traffic flow optimization for multiple intersections

Nam Bui, Khac-Hoai,Jung, Jason J. Elsevier 2018 Computers & electrical engineering Vol.71 No.-

<P><B>Abstract</B></P> <P>In this study, we focus on optimizing traffic flow at multiple intersections. Particularly, with the development of Internet of Things, intersection controllers are regarded as smart agents which can communicate and coordinate with each other. In this regard, a cooperative game theoretic approach among agents is proposed to improve traffic flow with large network. Thereby, a distributed merge and split algorithm for coalition formation is presented. This algorithm is applied to find out how to incorporate with the cooperation among agents for dynamically controlling traffic light at intersections. Furthermore, we construct a traffic simulation framework to evaluate our approach. With various parameters for traffic density, our proposed system can effectively improve traffic flow in both uniform and non-uniform. In particular, by coordinating among controllers, the waiting time of vehicles at intersections can be reduced from 15% to 25% comparing with previous methods (e.g., Green Wave Coordination).</P>

Computational negotiation-based edge analytics for smart objects

Bui, Khac-Hoai Nam,Jung, Jason J. Elsevier science 2019 Information sciences Vol.480 No.-

<P><B>Abstract</B></P> <P>In this paper, we propose a computational negotiation approach on Internet of Things (IoT) system where distributed edge devices can make their own decisions for smart applications. Particularly, Artificial Intelligence (AI) techniques play an important role of edge analytics in order to adaptively improve the performance of IoT systems. In this regard, we apply several AI techniques to provide negotiation models (e.g., synchronization, competition, and cooperation) among connected objects for edge analytics. Moreover, in the context of smart city, two typical use cases on IoT applications have been presented to evaluate our proposed approach. They are <I>i</I>) smart traffic control and <I>ii</I>) smart home energy management system.</P> <P><B>Highlights</B></P> <P> <UL> <LI> A distributed framework for communication and collaboration among connected objects is presented in this paper which can be applied for many critical issues. </LI> <LI> We apply computational negotiation models for connected objects which are expected to be an important issue for edge intelligence. </LI> <LI> We take two emergent IoT use cases for smart cities to evaluate the performances of our approach. </LI> </UL> </P>

Human Detection in Video Using Poselet with Articulated Pose Estimation and Edge-based RPCA Foreground Extraction

Ngoc Nam Bui,So Hee Min,Jin Young Kim 한국정보기술학회 2014 한국정보기술학회논문지 Vol.12 No.3

Human detection is an essential issue in image processing applications such as safety monitoring, interactive game, robot control, etc. However, most of current approaches focus on detecting pedestrians. In this paper, motivated by the high achievements of poselet technique, we propose a combinatorial approach between poselet technique and articulated pose estimation method to provide a robust hybrid structure for human detection in various environments. Also, a novel foreground extraction method, named edge based RPCA, is introduced to handle with the uncertain environments such as high illumination change or camera shaking problem. Then the candidate containing small foreground regions are discarded. In addition, an updating approach for the current human boundaries between consecutive frames are applied to prevent the abrupt change. To verify our proposed approach, the human detection experiments are conducted on the self-recorded database for both indoor and outdoor areas.

A Channel Fusion Approach Based GMM-UBM Supervector Using SVM with Non-Linear GMM KL and GUMI for Human Action Recognition

Ngoc Nam Bui,Thuong Khanh Tran,So Hee Min(민소희),Jin Young Kim(김진영) 한국정보기술학회 2015 한국정보기술학회논문지 Vol.13 No.3

Human Action Recognition (HAR), in recent years, has attracted much attention from the research community due to its challenges as well as wide applications. In this paper, we investigate Universal Background Model (UBM) based GMM supervector and Support Vector Machine (SVM) with dense trajectories and motion bound features for HAR system. A GMM supervector is obtained by MAP adaptation with UBM and cascading all the mean vector components. After that, supervectors are applied as input features to SVM classifier with several kernels including modified non-linear GMM KL and GUMI kernels. Moreover, we also adopted channel fusion that used to enhance the robustness of classify model. Then we make a comparison and critical analysis between our method with those existing systems. Experimental results demonstrates that the proposed approach performs more efficient than current systems.